1. Field of the Invention
The present invention relates to a technology for arranging orientation detection information for detecting an image orientation in an image.
2. Description of the Related Art
Conventionally, a technology called steganography with which an original image (primary information) is provided with additional information (secondary information) embedded therein that is invisible with human eye and the secondary information is retrieved on the image acquisition side.
For instance, an image processing apparatus disclosed in Japanese Patent Application Laid-Open No. 2004-349879 performs a process on an encoder side in which original image data is divided into several blocks and secondary information is embedded therein in accordance with the magnitude relation of the feature indices of individual blocks (for example, average density, granularity, color saturation, or centralization/decentralization of density). On the other hand, on a decoder, after the image data is divided again into blocks, the secondary information is retrieved based on the magnitude relation of the feature indices of the blocks.
In the above conventional technology, however, there is a problem that an image may not be processed in the correct orientation, as explained below.
Specifically, in the above conventional technology, there is a problem that, if the orientation of an image at the time of embedding secondary information in the original image data (at the time of encoding) is different from the orientation of the image at the time of retrieving the secondary information from the image data in which the secondary information is embedded (at the time of decoding), the image processing apparatus cannot accurately extract the magnitude relation of the feature indices residing in different blocks, which prevents the image from being processed in the correct direction.
In order to solve such a problem, there is a method in which information for detecting the orientation of original image data is embedded in the original image data. In particular, when the image processing apparatus encodes the original image data, information for detecting the orientation of the original image data (orientation detection information) is embedded in addition to the secondary information so that the image orientation can be detected by way of this orientation detection information at the time of decoding to process the image. The embedment of the orientation detection information here means to define, for instance, the magnitude relation of the feature indices indicating “top” and add modifications to the feature indices of blocks in the “top” region of the original image data in accordance with this definition.
However, according to this method, when the magnitude relation of the feature indices residing in blocks other than the orientation detection information coincides with the magnitude relation of the feature indices residing in the blocks for the orientation detection information indicating “top”, the image processing apparatus may mistake blocks other than the orientation detection information for the orientation detection information blocks, depending on the locations of the non-orientation detection information blocks, and make an erroneous judgment on the orientation of the image. Thus, there is a problem that the magnitude relation of the feature indices residing in the blocks cannot be correctly extracted, which prevents the image from being processed in the correct direction.
The above problem is not limited to the steganography but may arise in other types of image processing apparatuses (optical character reader (OCR), for instance) that do not employ the magnitude relation of the feature indices.